Cloudbeds has released a comprehensive study examining how generative AI platforms recommend hotels to travelers. Entitled “The Signals Behind Hotel AI Recommendations,” the study analyzed 145 consistently top-ranked properties across six global destinations to uncover the key factors driving AI hotel recommendations, according to the San Diego, Calif.-based AI-powered platform.
"The rules of travel discovery are being rewritten,” Adam Harris, CEO and co-founder of Cloudbeds, said in a statement. “Travelers no longer scroll through search results; they ask AI and get one answer. This report is the hospitality industry’s first in-depth look at how those answers are made. At Cloudbeds, we’re sharing these insights to help hoteliers lead in this new era of travel discovery."
The study was conducted using AI monitoring tools across ChatGPT, Perplexity and Gemini, which are platforms that collectively account for an estimated 98 percent of AI-driven website traffic, according to the company. Researchers analyzed citation patterns, sentiment analysis and digital footprints of consistently top-performing properties to identify the key factors influencing AI recommendations.
The study identified five strategies for hotels to secure visibility: curating a strategic OTA portfolio; optimizing hotel websites for accuracy and comprehensive detail; prioritizing reputation management across review platforms; expanding the digital footprint; and emphasizing unique brand storytelling.
The survey, meanwhile, recognized OTAs as dominant AI sources representing some 55 percent of all sources cited with TripAdvisor, Booking.com and Expedia leading the way. In addition, branded properties demonstrated a clear advantage as better than 72 percent of AI-recommended hotels were branded or large groups.
Furthermore, all recommended properties maintained excellent guest ratings and high review volumes across major platforms with an average sentiment score of 75 out of 100. Finally, the importance of a digital presence was reinforced as 98 percent of recommended properties appeared on YouTube; 97 percent in travel blogs; and 95 percent on Reddit.